Isolation, modification, and aldose reductase inhibitory activity of rosmarinic acid derivatives from the roots of Salvia grandifolia

Article abstract of DOI:10.1016/j.fitote.2016.05.011

To find aldose reductase inhibitors, two previously unreported compounds, grandifolias H and I, and five known compounds, including rosmarinic acid and rosmarinic acid derivatives, were isolated from the roots of Salvia grandifolia. A series of rosmarinic

Full text of DOI:10.1016/j.fitote.2016.05.011

Isolation, modification, and aldose reductase inhibitory activity of rosmarinic
acid derivatives from the roots of Salvia grandifolia
Jie Kang, Yanbo Tang, Quan Liu, Nan Guo, Jian Zhang, Zhiyan Xiao,
Ruoyun Chen, Zhufang Shen
PII:
DOI:
Reference:
S0367-326X(16)30120-4
doi: 10.1016/j.fitote.2016.05.011
FITOTE 3415
To appear in:
Fitoterapia
Received date:
Revised date:
Accepted date:
11 March 2016
20 May 2016
23 May 2016
Please cite this article as: Jie Kang, Yanbo Tang, Quan Liu, Nan Guo, Jian Zhang,
Zhiyan Xiao, Ruoyun Chen, Zhufang Shen, Isolation, modification, and aldose reductase
inhibitory activity of rosmarinic acid derivatives from the roots of Salvia grandifolia,
Fitoterapia (2016), doi: 10.1016/j.fitote.2016.05.011
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

ACCEPTED MANUSCRIPT
Isolation, modification, and aldose reductase inhibitory activity of rosmarinic
acid derivatives from the roots of Salvia grandifolia
a
a
a
a
b
a
Jie Kang , Yanbo Tang , Quan Liu , Nan Guo , Jian Zhang , Zhiyan Xiao , Ruoyun
a
a
Chen * and Zhufang Shen *
a
State Key Laboratory of Bioactive Substance and Function of Natural Medicines,
Institute of Materia Medica, Chinese Academy of Medical Sciences & Peking Union
Medical College, Beijing 100050, People’s Republic of China
b
Department of Cell and Molecular Biology, Research Institute of Orthopedics &
Traumatology, Foshan Hospital of TCM, Foshan 528000, People’s Republic of China
Abbreviated running title: Aldose reductase inhibitors from S. grandifolia
*
Corresponding authors. Tel/Fax: +86-10-8316-1622 (R.C.), +86-10-8317-2669
(Z.S.). E-Mail: rych@imm.ac.cn (R.C.), shenzhf@imm.ac.cn (Z.S.).
1

ACCEPTED MANUSCRIPT
ABSTRACT
To find aldose reductase inhibitors, two previously unreported compounds,
grandifolias H and I, and five known compounds, including rosmarinic acid and
rosmarinic acid derivatives, were isolated from the roots of Salvia grandifolia. A
series of rosmarinic acid derivatives was obtained from rosmarinic acid using simple
synthetic methods. The aldose reductase inhibitory activity of the isolated and
synthesized compounds was assessed. Seven of the tested compounds showed
moderate aldose reductase inhibition (IC = 0.06–0.30 μM). The structure-activity
5
0
relationship of aldose reductase inhibitory activity of rosmarinic acid derivatives was
discussed for the first time. This study provided useful information that will facilitate
the development of aldose reductase inhibitors.
2

ACCEPTED MANUSCRIPT
Keywords
Salvia grandifolia; isolation; modification; grandifolias H and I; rosmarinic acid
derivative; aldose reductase inhibition
3

ACCEPTED MANUSCRIPT
1
. Introduction
Diabetes is a chronic disease characterized by an impaired response to insulin
and/or progressively reduced function of pancreatic -cells [1]. All forms of diabetes
are characterized by chronic hyperglycemia and the development of diabetes-specific
microvascular pathology in the retina, renal glomerulus, and peripheral nerves, in
addition to other tissues. Diabetes is also associated with accelerated atherosclerotic
macrovascular disease, which affects arteries that supply the heart, brain, and lower
extremities [2,3]. Diabetes and its complications are a major cause of morbidity and
mortality worldwide and currently affect almost 400 million people. The number of
patients afflicted with type 2 diabetes is increasing rapidly in developed and
developing countries [1,4].
Aldose reductase is a central enzyme in the polyol pathway that belongs to the
aldo-keto reductase superfamily and is implicated in aberrant glucose metabolism and
diabetic complications [4-6]. Several studies have demonstrated the important role of
aldose reductase in accelerating atherosclerosis and vascular injury associated with
diabetes and aging. In hyperglycaemic conditions, aldose reductase catalyses
NADPH-dependent reduction of glucose to sorbitol, which in turn is oxidized to
+
fructose by an NAD -dependent sorbitol dehydrogenase. Once sorbitol is accumulated
inside the cells, it can not diffuse easily across the cell membrane as a result osmotic
pressure increases causing cellular damage. Aldose reductase inhibitors have been
shown to reduce tissue sorbitol accumulation in diabetic animals and there are
evidences that blockage of aldose reductase can have beneficial effect in diabetic
4

ACCEPTED MANUSCRIPT
complications [7-8]. Up to now, various small molecule aldose reductase inhibitors
have been evaluated in preclinical and clinical trials [9-12]. However, epalrestat is the
only commercially available aldose reductase inhibitor in Japan in 2015 [13]. Hence,
there is an urgent need for new and effective aldose reductase inhibitors.
Salvia, the largest genus in the family Lamiaceae, comprises over 900 species
that are distributed globally [14-16]. The genus Salvia has been assessed in many
studies because it is a rich source of polyphenol compounds, of which more than 160
have been isolated from plants in the genus, some of which are unique to Salvia [15].
The majority of polyphenolics isolated from Salvia probably originate from
condensation events on caffeic acid, their common building block [17,18]. The roots
of Salvia grandifolia, a plant endemic to China, have been used to treat cardiovascular
diseases in Sichuan and Yunnan Provinces of China for centuries.
Rosmarinic acid is an ester of caffeic acid and 3,4-dihydroxyphenyllactic acid
that is found in a variety of plants, but it is especially abundant in the subfamily
Nepetoideae of the Lamiaceae [19]. Rosmarinic acid has been identified in more than
7
7 species of Lamiaceae at a concentration ranging from 0.01 to 9.30 mg rosmarinic
acid per gram of plant material [20,21]. Rosmarinic acid has several biological
activities with potential therapeutic applications, including antiviral, antibacterial,
anti-inflammatory, antioxidant, and aldose reductase inhibitory effects [22-24].
To our knowledge, there have been no reports on water soluble polyphenol
compounds in S. grandifolia. We found that the EtOAc extract of S. grandifolia was
rich in rosmarinic acid and showed good aldose reductase inhibition (65% at 10
5

ACCEPTED MANUSCRIPT
μg/mL), which inspired us to further study hydrosoluble polyphenols in the extract,
especially rosmarinic acid derivatives. According to the functional groups of
rosmarinic acid, semi-synthetic compounds were prepared. All of the rosmarinic acid
derivatives both from nature and synthesis were investigated the relationship between
the structures and aldose reductase inhibitory effects.
2
. Experimental part
2
.1. General
Optical rotation was measured with a Jasco P-2000 polarimeter (Tokyo, Japan).
UV spectra were collected in MeOH on a Jasco V-650 spectrophotometer (Tokyo,
Japan). IR spectra were recorded on a Nicolet 5700 spectrometer (Madison, WI,
1
13
USA) by the FT-IR transmission electron microscopy method. H and C NMR
spectra were acquired using a Varian MP-400 spectrometer (Palo Alto, CA, USA) or
an INOVA (500 or 600 M) spectrometer (Varian, Palo Alto, CA, USA). HRESIMS
were recorded on an Agilent 1200 SL series LC/6520 QTOF spectrometer
(Boeblingen, Germany). Column chromatography (CC) purification was performed
using silica gel (160–200 mesh, Qingdao Marine Chemical Factory, Qingdao, China),
Sephadex LH-20 (GE Healthcare Bio-Sciences AB, Uppsala, Sweden), and C-18 (50
μm, YMC, Kyoto, Japan). CC fractions were analyzed by TLC (silica gel GF254,
Qingdao Marine Chemical Factory, Qingdao, China). A μQuant MQX200 microplate
6

ACCEPTED MANUSCRIPT
reader (BIO-TEK, USA) was used to assay aldose reductase inhibitory activity.
NADPH-Na , D,L-glyceraldehyde, and epalrestat were obtained from Sigma-Aldrich
4
(Shanghai) Trading Co., Ltd. (Shanghai, China). All analytical grade solvents used for
isolation and separation were obtained from Beijing Chemical Factory (Beijing,
China).
2
.2. Plant materials
Roots of Salvia grandifolia W. W. Smith (Liabiatae) were collected from Yunnan
Province, China, in October 2007, and identified by Professor Lin Ma of the Institute
of Materia Medica, Chinese Academy of Medical Science and Peking Union Medical
College, China. A voucher specimen (ID-21587) was deposited in the Institute of
Materia Medica, Chinese Academy of Medical Science and Peking Union Medical
College, China.
2
.3. Extraction and isolation
Air dried, powdered roots of S. grandifolia (13.3 kg) were extracted with
EtOH-H O (95:5, v/v, 3, 150 L each time) under reflux for 2 h. Evaporation of the
2
solvent under reduced pressure gave a residue (1.05 kg), which was suspended in H O
2
and successively partitioned with CHCl , EtOAc, and n-BuOH, respectively. The
3
EtOAc extract (290 g) was applied to a silica gel column (160–200 mesh, 2.0 kg) and
subjected to gradient elution (petroleum ether : acetone = 9:1–1:1, v/v) to yield nine
7

ACCEPTED MANUSCRIPT
fractions (A–I). Fraction D (25.4 g) was subjected to C-18 CC (50 μm, 800 g) and
eluted in a gradient eluent (MeOH:H O = 3:7–9:1, v/v) to give 1 (10 mg), 4 (3.1 g),
2
and 5 (2.9 g). Fraction F (24.3 g) was applied to a C-18 column (50 μm, 800 g) and
eluted with an isocratic eluent (MeOH:H O = 2:8, v/v) to give 3 (4.2 g) and 6 (11
2
mg). Fraction H (19.0 g) was applied to a C-18 column (50 μm, 500 g) and eluted
with an isocratic eluent (MeOH:H O = 6:4, v/v) to give seven subfractions, H1–H7.
2
Subfraction H4 was applied to a Sephadex LH-20 column and eluted with
MeOH:H O (6:4, v/v) to yield 2 (11 mg) and 7 (12 mg).
2
2
.4. Absolute sugar configuration
The absolute configuration of glucose was determined according to a reported
procedure [25]. Compound 2 (2.2 mg) was hydrolyzed with 1 M HCl (1 mL) at
00 °C for 2 h, diluted with H O, and extracted with EtOAc (3  5 mL). After the
1
2
H O layer was dried under vacuum, the residue was dissolved in pyridine (0.5 mL)
2
containing L-cysteine methyl ester hydrochloride (2 mg) and heated at 60 C for 2 h.
o-Tolylisothiocyanate (2 μL) was added and the mixture was heated at 60 C for 2 h.
The reaction mixture was directly analyzed by an Agilent 1260 HPLC with a DAD
detector at 254 nm. A Cosmosil 5C18-AR-II HPLC column (5 μm, 150 mm  4.6 mm,
Nacalai Tesque Inc., Nakagyo-ku, Japan) was used at 35 C. Each sample was eluted
in an isocratic step of CH CN-H O (25:75) containing 50 mM H PO with a flow rate
3
2
3
4
of 0.8 mL/min. The HPLC column was washed with MeOH after each injection. The
8

ACCEPTED MANUSCRIPT
reaction conditions described above were also used for the analysis of D- and
L-glucose. The retention times of D-glucose (12.65 min) and L-glucose (11.62 min)
were compared with those of the reaction mixtures. The sugar derivatives from 2
showed a peak at 12.56 min, which coincided with derivatives of D-glucose.
1
13
2
.5. HRESIMS, H and C NMR spectral data of compounds 1–2
Grandifolia H (1)
25
Red amorphous powder; [α]
-119 (c 0.11, MeOH); UV (MeOH) max (log ε)
D
2
05 (2.83) nm, 219 (2.73) nm, 251 (2.54) nm, 274 (2.60) nm, 339 (2.32) nm; ECD (c,
-3
1
10 M, MeOH)  ( ) 330 (-0.02), 367 (+0.03), 435 (+0.02), 501 (-0.04); IR ν
max
1
13
3
432, 2939, 1740, 1670, 1602, 1284; for H NMR (MeOD, 500 MHz) and C
max
NMR (MeOD, 125 MHz) spectroscopic data, see Tables 1 and 2; HRESIMS (negative
-
ion) m/z 653.1659 [M - H] (calcd. for C H O , 653.1664).
3
6
29 12
Grandifolia I (2)
2
5
White amorphous powder; [α] -30 (c 0.09, MeOH); UV (MeOH) max (log ε)
D
2
3
06 (2.34) nm, 228 (2.41) nm, 268 (2.19) nm, 291 (2.36) nm, 318 (2.28) nm; IR ν max
1
13
357, 2974, 1694, 1601, 1516, 1285; for H NMR (MeOD, 600 MHz) and C NMR
(MeOD, 150 MHz) spectroscopic data, see Tables 1 and 2; HRESIMS (negative ion)
-
m/z 671.1621 [M - H] (calcd. for C H O , 671.1618).
3
2
31 16
9

ACCEPTED MANUSCRIPT
2
2
2
.6. Synthesis
.6.1. Synthesis of compounds 8-17
.6.1.1. Preparation of compound 8
A stirred solution of 3 (180 mg, 0.5 mmol) in DMF (5 mL) was added to
anhydrous K CO (520 mg, 3.75 mmol) followed by methyl iodide (0.24 mL, 3.75
2
3
mmol). The reaction mixture was heated to 60 ºC for 2 d, after which methyl iodide
0.24 mL, 3.75 mmol) was added and heating was continued for a further 2 d. The
(
mixture was cooled to room temperature and quenched by the addition of H O (30
2
mL), after which the resulting mixture was extracted with Et O (3  40 mL) and the
2
combined organic phases were dried over anhydrous Na SO and concentrated in
2
4
vacuo. The residue was purified by silica gel column chromatography
EtOAc:petroleum ether = 2:5) to afford 8 as a colorless oil (160 mg, 74% yield).
(
2
.6.1.2. Preparation of compounds 12 and 13
Rosmarinic acid (3) (300 mg) was resolved in acetone (5 mL), after which 5%
palladium on activated carbon (40 mg) was added to the solution. The reaction
mixture was stirred under a hydrogen atmosphere at room temperature for 3 h and
filtered through celite, affording 12 as colorless oil (296 mg, 98% yield).
10

ACCEPTED MANUSCRIPT
Using compound 8 as a starting material, compound 13 was obtained as a
colorless oil (95% yield) by the same procedure.
2
.6.1.3. Preparation of compounds 9–11 and 14–17
A mixture of rosmarinic acid (3) or 7,8-saturated rosmarinic acid (12) (1 equiv.)
was stirred overnight at room temperature with different amine substituents (1.2
equiv.), N,N-Diisopropylethylamine (DIEA, 2.5 equiv.), and
-(7-aza-1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate
HATU, 1.2 equiv.) in CH Cl and concentrated in vacuo. The residue was resolved in
2
(
2
2
EtOAc and washed with H O. The organic phase was separated and concentrated in
2
vacuo to give a residue, which was purified by silica gel column chromatography
(
CH Cl :CH OH:CH COOH = 45:1:0.02) to afford target compounds (9–11 and
2 2 3 3
1
4–17, the yields of all amino compounds about 85%-95%).
1
13
2
.6.2. HRESIMS, H, and C NMR spectral data of compounds 8–17
+
Compound 8, HRESIMS (positive ion) m/z 453.1525 [M+Na] (calcd. for
1
C H O Na, 453.1520). H NMR (400 MHz, DMSO-d ) : 7.62 (1H, d, J = 16.0 Hz),
23
26
8
6
7
.36 (1H, d, J = 2.0 Hz), 7.25 (1H, dd, J =8.0, 2.0 Hz), 6.99 (1H, d, J = 8.0 Hz), 6.91
1H, d, J = 2.0 Hz), 6.87 (1H, d, J = 8.0 Hz), 6.80 (1H, dd, J = 8.0, 2.0 Hz), 6.57 (1H,
d, J = 16.0 Hz), 5.25 (1H, m), 3.81 (3H, s), 3.80 (3H, s), 3.73 (3H, s), 3.71 (3H, s),
(
11

ACCEPTED MANUSCRIPT
1
3
3
1
.67 (3H, s), 3.10 (2H, m). C NMR (100 MHz, DMSO-d ) : 169.9, 165.8, 151.2,
6
49.0, 148.4, 147.7, 145.9, 128.3, 126.6, 123.3, 121.3, 114.4, 113.0, 111.6, 111.5,
110.4, 72.7, 55.6, 55.5, 55.4, 55.3, 52.0, 36.3.
+
Compound 9, HRESIMS (positive ion) m/z 450.1543 [M+H] (calcd. for
1
C H NO , 450.1547). H NMR (400 MHz, DMSO-d ) : 9.62 (1H, s, OH), 9.15 (1H,
25
24
7
6
s, OH), 8.75 (1H, s, OH), 8.71 (1H, s, OH), 8.57 (1H, br s, NH), 7.46 (1H, d, J = 15.6
Hz), 7.21 (4H, m), 7.00 (3H, m), 6.76 (1H, m), 6.63 (2H, m), 6.51 (1H, m), 6.25 (1H,
1
3
d, J = 15.6 Hz), 5.15 (1H, br s), 4.29 (2H, br s), 2.90 (2H, m). C NMR (100 MHz,
DMSO-d ) : 169.0, 165.8, 148.5, 145.6 (2), 144.9, 143.9, 139.1, 128.2 (2), 127.4,
6
1
3
26.9 (2), 126.6, 125.4, 121.5, 120.1, 116.8, 115.7, 115.3, 114.8, 113.6, 74.3, 41.8,
6.9.
+
Compound 10, HRESIMS (positive ion) m/z 480.1649 [M+H] (calcd. for
1
C H NO , 480.1653). H NMR (400 MHz, DMSO-d ) : 9.63 (1H, s, OH), 9.17 (1H,
26
26
8
6
s, OH), 8.77 (1H, s, OH), 8.72 (1H, s, OH), 8.51 (1H, br s, NH), 7.47 (1H, d, J = 16.0
Hz), 7.08 (4H, m), 6.86 (3H, m), 6.65 (2H, m), 6.50 (1H, d, J = 7.2 Hz), 6.25 (1H, d, J
1
3
=
16.0 Hz), 5.15 (1H, br s), 4.29 (2H, br s), 3.74 (3H, s), 2.90 (2H, m). C NMR (100
MHz, DMSO-d ) : 168.9, 165.8, 158.1, 148.5, 145.6 (2), 144.9, 143.9, 131.0, 128.3
6
(2), 127.4, 125.4, 121.5 (2), 120.0, 116.7, 115.7, 115.3, 114.8, 113.6 (2), 74.2, 55.0,
4
1.2, 36.9.
12

ACCEPTED MANUSCRIPT
+
Compound 11, HRESIMS (positive ion) m/z 464.1704 [M+H] (calcd. for
1
C H NO , 464.1715). H NMR (400 MHz, DMSO-d ) : 9.62 (1H, s, OH), 9.16 (1H,
26
26
7
6
s, OH), 8.75 (1H, s, OH), 8.71 (1H, s, OH), 8.49 (1H, d, J = 4.4 Hz, NH), 7.46 (1H, d,
J = 16.0 Hz), 7.28 (2H, m), 7.22 (3H, m), 7.05 (1H, s), 7.00 (1H, d, J = 7.6 Hz), 6.77
(1H, d, J = 8.0 Hz), 6.69 (1H, s), 6.62 (1H, d, J = 8.0 Hz), 6.49 (1H, d, J = 7.6 Hz),
6
6
1
.25 (1H, d, J = 16.0 Hz), 4.94 (1H, m), 4.55 (1H, m), 2.84 (2H, m), 1.36 (3H, d, J =
1
3
.8 Hz). C NMR (100 MHz, DMSO-d ) : 168.1, 165.9 148.5, 145.6, 145.5, 144.9,
6
44.2, 143.9, 128.2 (2), 127.4, 126.5, 125.8 (2), 125.4, 121.5, 120.0, 116.7, 115.7,
115.3, 114.8, 113.6, 74.2, 47.6, 36.9, 22.2.
+
Compound 12, HRESIMS (positive ion) m/z 385.0893 [M+Na] (calcd. for
1
C H O Na, 385.0894). H NMR (400 MHz, DMSO-d ) : 12.88 (1H, br s, COOH),
18
18
8
6
8
6
.79 (1H, s, OH), 8.70 (2H, s, OH), 8.63 (1H, s, OH), 6.62 (2H, m), 6.56 (2H, m),
.46 (1H, dd, J = 8.0, 1.6 Hz), 6.36 (1H, dd, J = 8.0, 1.6 Hz), 4.93 (1H, m), 2.83 (2H,
1
3
m), 2.62 (2H, m), 2.51 (2H, m). C NMR (100 MHz, DMSO-d ) : 171.7, 170.6,
6
1
3
45.0, 144.9, 144.0, 143.5, 131.1, 127.0, 120.0, 118.7, 116.6, 115.6, 115.4, 115.3, 72.8,
5.9, 35.4, 29.5.
+
Compound 13, HRESIMS (positive ion) m/z 455.1681 [M+Na] (calcd. for
1
C H O Na, 455.1676). H NMR (400 MHz, DMSO-d ) : 6.86 (2H, m), 6.80 (2H,
23
28
8
6
m), 6.72 (1H, dd, J = 8.0, 1.6 Hz), 6.62 (1H, dd, J = 8.0, 1.6 Hz), 5.13 (1H, m), 3.73
3H, s), 3.72 (3H, s), 3.71 (6H, s), 3.64 (3H, s), 3.01 (2H, m), 2.76 (2H, m), 2.63 (2H,
(
13

ACCEPTED MANUSCRIPT
1
3
m). C NMR (100 MHz, DMSO-d ) : 171.7, 169.6, 148.6, 148.4, 147.6, 147.2,
6
1
3
32.6,128.1, 121.3, 119.8, 113.0, 112.1, 111.8, 111.6, 72.7, 55.5, 55.4, 55.3 (2), 52.0,
6.1, 35.0, 29.6.
+
Compound 14, HRESIMS (positive ion) m/z 452.1714 [M+H] (calcd. for
1
C H NO , 452.1704). H NMR (400 MHz, DMSO-d ) : 8.78 (1H, s, OH), 8.70 (2H,
25
26
7
6
s, OH), 8.65 (1H, s, OH), 8.46 (1H, m, NH), 7.31 (2H, m), 7.22 (1H, m), 7.11 (2H, d,
J = 7.2 Hz), 6.60 (4H, m), 6.45 (1H, dd, J = 8.0, 2.0 Hz), 6.36 (1H, dd, J = 8.0, 2.0
1
3
Hz), 5.06 (1H, m), 4.27 (2H, m), 2.85 (2H, m), 2.58 (4H, m). C NMR (100 MHz,
DMSO-d ) : 171.7, 168.8, 145.0, 144.9, 143.9, 143.4, 139.0, 131.1, 128.2 (2), 127.2,
6
1
26.9 (2), 126.6, 120.1, 118.7, 116.8, 115.6, 115.4, 115.3, 74.2, 41.7, 36.8, 35.4, 29.5.
+
Compound 15, HRESIMS (positive ion) m/z 482.1807 [M+H] (calcd. for
1
C H NO , 482.1809). H NMR (400 MHz, DMSO-d ) : 8.70 (4H, br s, OH), 8.35
26
28
8
6
(1H, br s, NH), 7.00 (2H, d, J = 8.0 Hz), 6.82 (2H, d, J = 8.0 Hz), 6.58 (4H, m), 6.41
(1H, d, J = 8.0 Hz), 6.33 (1H, d, J = 8.0 Hz), 5.03 (1H, m), 4.16 (2H, m), 3.70 (3H, s),
1
3
2
1
.80 (2H, m), 2.54 (4H, m). C NMR (100 MHz, DMSO-d ) : 171.6, 168.6, 158.1,
6
45.0, 144.9, 143.9, 143.4, 131.1, 131.0 128.2 (2), 127.2, 120.1, 118.7, 116.8, 115.6,
115.4, 115.3, 113.6 (2), 74.2, 55.0, 41.2, 36.8, 35.5, 29.5.
+
Compound 16, HRESIMS (positive ion) m/z 466.1864 [M+H] (calcd. for
1
C H NO , 466.1860). H NMR (400 MHz, DMSO-d ) : 8.76 (1H, s, OH), 8.71 (1H,
26
28
7
6
14

ACCEPTED MANUSCRIPT
s, OH), 8.70 (1H, s, OH), 8.64 (1H, s, OH), 8.38 (1H, d, J = 7.6 Hz, NH), 7.29 (2H,
m), 7.20 (3H, m), 6.61 (4H, m), 6.44 (1H, dd, J = 8.0, 2.0 Hz), 6.36 (1H, dd, J = 8.0,
2
.0 Hz), 5.04 (1H, m), 4.91 (1H, m), 2.85 (1H, m), 2.83 (1H, m), 2.60 (2H, m), 2.51
1
3
(
2H, overlapped), 1.34 (3H, d, J = 6.8 Hz). C NMR (100 MHz, DMSO-d ) : 171.6,
6
1
1
67.9, 145.0, 144.9, 144.2, 143.9, 143.4, 131.1, 128.1 (2), 127.2, 126.5, 125.8 (2),
20.0, 118.7, 116.7, 115.6, 115.4, 115.3, 74.3, 47.6, 36.8, 35.5, 29.6, 22.2.
+
Compound 17, HRESIMS (positive ion) m/z 524.1923 [M+H] (calcd. for
1
C H NO , 524.1915). H NMR (400 MHz, DMSO-d ) : 8.73 (1H, s, OH), 8.69 (1H,
28
30
9
6
s, OH), 8.65 (1H, s, OH), 8.61 (1H, s, OH), 8.47 (1H, d, J = 8.0 Hz, NH), 7.20 (4H,
m), 6.56 (4H, m), 6.33 (3H, m), 4.92 (1H, m), 4.46 (1H, m), 3.62 (3H, s), 3.03 (1H,
1
3
m), 2.89 (1H, m), 2.65 (6H, m, overlapped). C NMR (100 MHz, DMSO-d ) : 171.6,
6
1
1
71.3, 168.9, 145.0, 144.9, 143.9, 143.4, 137.1, 131.1, 129.1 (2), 128.1 (2), 127.4,
26.5, 119.9, 118.6, 116.6, 115.5, 115.4, 115.3, 73.8, 55.8, 53.3, 36.8, 36.5, 35.4, 29.6.
2
.7. Aldose reductase inhibitory assay
Measurements were based on a previously reported method [26]. Crude aldose
reductase was prepared from rat lens. Two lenses were homogenized in 1 mL of 100
mM sodium phosphate buffer at pH 6.2. The homogenate was centrifuged at 15000 ×
g for 30 min at 4 C and the resulting supernatant was used as the source of aldose
reductase. The assay mixture contained 10 mM D,L-glyceraldehyde and 0.16 mM
NADPH-Na in 1 mL of 50 mM sodium phosphate buffer (pH 6.2), with or without
4
15

ACCEPTED MANUSCRIPT
the test compound and enzyme preparation (rat lens enzyme). Appropriate blanks that
contains sodium phosphate buffer and enzyme preparation were employed for
corrections. The assay mixture was incubated at 25 C. After 10 min of incubation,
the plate was immediately cooled at -20 °C for 5 min to stop the reaction. The change
in the absorbance at 340 nm due to NADPH oxidation was measured in a plate reader.
All data are given as the mean ± standard deviations.
3
. Results and discussion
3
.1. Structural analysis of compounds
The EtOAc fraction of the 95% EtOH extract of the roots of S. grandifolia was
subjected to column chromatography over silica gel, C-18, and Sephadex LH-20,
respectively, with purification using preparative HPLC to yield two new and five
known rosmarinic acid derivatives (including rosmarinic acid) (Fig. 1). The known
compounds were identified through a comparison of their NMR spectroscopic data
with those reported in the literature, and they were totally matched: rosmarinic acid
(3) [27], methyl rosmarinate (4) [28], ethyl rosmarinate (5) [28],
3
-O-methyl-rosmarinic acid (6) [28], salviaflaside (7) [27].
Compound 1, had the molecular formula C H O , as established by HRESIMS
3
6
30 12
-
[
M - H] m/z 653.1659, which indicated 22 degrees of unsaturation. The IR spectrum
16

ACCEPTED MANUSCRIPT
-1
-1
of 1 indicated the presence of hydroxyl (3432 cm ), carbonyl (1740, 1670 cm ), and
-1
1
phenyl ring (1602 cm ) functionalities. The H NMR spectrum showed resonances
(
Table 1) for two methyl (δ 2.19, H -17; 1.43, H -18), three methylene (δ 3.13
H
3
3
H
(
2H, t), H-7; 2.95 (m), 2.60 (m), H-1; 1.91 (m), 1.84 (m), H-2), two oxymethine (δH
.23 (t), H-3; 5.17 (t), H-8), and 11 methine (δ 7.90 (d), H-6; 7.48 (d), H-7; 7.26
5
H
(d), H-15; 6.95 (d), H-7; 6.82 (d), H-5; 6.74 (d), H-2; 6.65 (dd), H-6; 6.52 (d), H-2;
1
3
6
.50 (dd), H-6; 6.48 (d), H-5; 5.61 (d), H-8) groups. The C NMR spectrum
displayed a total of 36 carbon resonances (Table 2), assignable to four ketocarbonyl
(
δ 184.8, C-11; 176.7, C-12; 171.3, C-9; 170.5, C-9), two methyl (δ 30.2, C-18;
C
C
9
.7, C-17), three methylene (δ 38.0, C-7; 25.7, C-2; 25.5, C-1), two oxymethine
C
(
δ 76.9, C-3; 75.9, C-8), 11 methine (δ 148.3, C-7; 143.9, C-15; 135.3, C-6;
C
C
1
24.2, C-6; 123.0, C-6; 122.5, C-7; 118.9, C-2; 117.1, C-5, 5; 115.3, C-2; 113.4,
C-8), and 14 quaternary carbon (δ 163.0, C-14; 150.7, C-4; 147.4, C-3; 147.1, C-5;
C
1
46.9, C-3; 146.5, C-4; 144.6, C-10; 131.1, C-8; 129.0, C-1; 127.5, C-1; 126.9,
1
13
C-9; 122.5, C-13; 121.5, C-16; 73.2, C-4) groups (Table 2). The H and C NMR
spectra of 1 showed rosmarinic acid and tanshinone parts in its structure. The HMBC
correlation of H-3/C-9 (Fig. 2) suggested an ester group formed by the hydroxyl
group of tanshinone and the carboxyl group of rosmarinic acid. Thus, the planar
structure of 1 was established as shown in Fig. 1. NOEs were observed between
H-3(5.23, t, J = 2.0 Hz) and CH -18 (1.43, 3H, s) suggested those protons on the
3
same side of the molecule (Fig. 2), which determined the structure of tanshinone part
is the same as that of tanshindiol C [29]. However, NOEs of H-8ʹ and H-3 were not
17

ACCEPTED MANUSCRIPT
shown. Because the absolute configuration of natural rosmarinic acid is R (C-8ʹ ), it
was concluded that the absolute configuration of C-8ʹ is R. Since the absolute
configuration of C-8ʹ was determined, the structure of compound 1 was symplified
to 1-A and 1-B (Fig. 3a) to make the calculation of the other chiral carbons easier.
Conformational analysis of 1-A (Fig. 3a) showed four lowest energy conformers
whose relative energy within 4 kcal/mol. The conformers were further optimized at
the B3LYP/6-31G(d) level. Rotational strengths of the 100 lowest electronic
transitions were calculated. ECD spectra of different conformers were simulated with
a half-bandwidth of 0.2 eV. The overall theoretical ECD spectra were then obtained
according to the Boltzmann weighting of each conformers. The ECD spectrum of 1
was recorded and showed positive (367, 435 nm) and negative (330, 501 nm) Cotton
effects, which was similar as the calculated ECD curve of 1-A (3S, 4R), but was
opposite to that of 1-B (3R, 4S) (Fig. 3a and b). Therefore the absolute
configurations of C-8ʹ , 3 and 4 were determined to be R, S and R. The structure of
1
(grandifolia H) was established as shown (Fig. 1). This is the first report of an ester
compound formed by tanshinone and rosmarinic acid.
Compound 2, obtained as a white amorphous powder, had the molecular formula
-
C H O , as determined by HRESIMS ([M - H] m/z 671.1621), which indicated 17
32
32 16
degrees of unsaturation. The 1D (Tables 1 and 2) and 2D NMR spectroscopic data of
2
were similar to that of known compound 7 [27], except for an extra
-methoxy-4-hydroxy-benzoyl group connecting the hydroxyl group at C-6 in 2,
3
which was further confirmed by the HMBC correlations of H-6, 2, and 6/C-7, as
18

ACCEPTED MANUSCRIPT
well as NOE difference experiments (Fig. 2). Irradiation of the methoxy group at 
3
.76 enhanced hydrogen signal H-2 at  7.44 (Fig. 2), which was indicative of the
methoxy group at C-3, but not C-4. The coupling constant (J = 6.5 Hz) of
anomeric hydrogen ( 4.95) suggested a -glucopyranosyl moiety. The identity of
D-glucose was confirmed by a reported procedure [25]. Because the absolute
configuration of natural rosmarinic acid is R (C-8ʹ ), and the optical rotation value of
2
is negative, which is similar as that of hedyotoside E [30], so the absolute
configuration of C-8ʹ was determined to be R. The structure of 2 (grandifolia I) was
established as shown (Fig. 1).
3
.2. Synthesis of compounds 8–17
To determine the aldose reductase inhibitory activity of different substituents of
rosmarinic acid derivatives, ten rosmarinic acid derivatives (8–17) (Fig.1) were
synthesized according to the method described in Fig. 4. Among the synthesized
rosmarinic acid derivatives, compounds 9, 10, and 14–17 have not been reported
previously. The absolute configurations of C-8ʹ of rosmarinic acid derivatives (8–17)
are R, due to no configurations changed during the synthetic process.
3
.3. Aldose reductase inhibitory activities of compounds 1–17
Aldose reductase is the first enzyme of the polyol pathway, which converts
19

ACCEPTED MANUSCRIPT
glucose into sorbitol and is therefore considered to be responsible for diabetic
complications [31]. Rosmarinic acid (3) inhibits aldose reductase at relatively low
concentration [32-34]. Herein, the structure-activity relationship of rosmarinic acid
derivatives with regard to aldose reductase inhibition was discussed for the first time.
The aldose reductase inhibitory activities of seven natural (1–7) and ten modified
(8–17) rosmarinic acid derivatives were evaluated by comparing their effects with that
of positive control epalrestat. Seven compounds (1, 3, 4, 5, 9, 10, and 11) had IC50
values within the range of 0.06–0.30 μM (Fig. 5, Table 3), indicating greater potency
than that of natural aldose reductase inhibitor quercetin (IC = 5.20 μM) [32]. The
5
0
other ten tested compounds showed very weak aldose reductase inhibitory activities
inhibition rates at a concentration of 10 μM were 5.7–24.2%). Compounds 2 and 7
are rosmarinic acid derivative glycosides with aldose reductase inhibitory rates of
8.6% and 26.9% (10 μM), respectively, suggesting that glycosylation attenuated
(
2
aldose reductase inhibition. 7,8-Saturated rosmarinic acid derivatives 12–17 also
showed weak aldose reductase inhibitory rates (9.3%–30.6%) at a concentration of 10
μM, indicating that the double bond of the α,-unsaturated ketone is the key moiety
for aldose reductase inhibition. The aldose reductase inhibitory rates of compounds 6
and 8, containing phenolic hydroxyl groups with partial (6) or total (8) methylation,
were 46.3% and 18.9%, respectively, at a concentration of 10 μM. These results
showed that the degree of methylation of the phenolic hydroxyl groups of rosmarinic
acid derivatives is inversely related to aldose reductase inhibitory activity.
The results described above suggest that the presence of a double bond in the
20

ACCEPTED MANUSCRIPT
α,-unsaturated ketone unit of rosmarinic acid derivatives is crucial for aldose
reductase inhibition. Glycosylation and methylation of phenolic hydroxyl groups of
rosmarinic acid derivatives were inactive.
4
. Conclusion
Two new rosmarinic acid derivatives grandifolias H and I (1-2) were isolated
from the roots of S. grandifolia, with a series of rosmarinic acid derivatives (8-17)
obtained from rosmarinic acid (3). To the best of our knowledge, this is the first report
of an ester compound (1) formed by tanshinone and rosmarinic acid. Bioactivity
experiments showed that seven rosmarinic acid derivatives (1, 3, 4, 5, 9, 10, and 11)
had moderate concentration-dependent aldose reductase inhibitory effects. The
structure-activity relationship of rosmarinic acid derivatives with regard to aldose
reductase inhibition was discussed for the first time. This study provided useful
information that will facilitate the development of aldose reductase inhibitors. Further
studies are warranted to reveal the mechanisms underlying the aldose reductase
inhibitory effects of rosmarinic acid derivatives.
Acknowledgements
We gratefully acknowledge financial support from the National Natural Science
Foundation of China (21302227) and the Scientific Research Foundation for Returned
Overseas Chinese Scholars, State Education Ministry, China.
21

ACCEPTED MANUSCRIPT
There are no conflicts of interest to declare.
22

ACCEPTED MANUSCRIPT
References
[
1] R. Maccari, R. Ottana, Targeting aldose reductase for the treatment of diabetes
complications and inflammatory diseases: New insights and future directions, J.
Med. Chem. 58 (2015) 2047-2067.
[
[
2] M. Brownlee, Biochemistry and molecular cell biology of diabetic complications,
Nature 414 (2001) 813-820.
3] X. Hu, S. Li, G. Yang, H. Liu, G. Boden, L. Li, Efficacy and safety of aldose
reductase inhibitor for the treatment of diabetic cardiovascular autonomic
neuropathy: Systematic review and meta-analysis, Plos One 9 (2014) e87096.
4] S. Vedantham, R. Ananthakrishnan, A. Schmidt, R. Ramasamy, Aldose reductase,
oxidative stress and diabetic cardiovascular complications, Cardiovasc. Hematol,
Agents Med. Chem. 10 (2012) 234-240.
[
[
[
[
[
5] J.M. Petrash, I. Tarle, D.K. Wilson, F.A. Quiocho, Aldose reductase catalysis and
crystallography. Insights from recent advances in enzyme structure and function,
Diabetes 43 (1994) 955-959.
6] C.E. Grimshaw, K.M. Bohren, C.J. Lai, K.H. Gabbay, Human aldose reductase:
rate constants for a mechanism including interconversion of ternary complexes
by recombinant wild-type enzyme, Biochemistry 34 (1995) 14356-14365.
7] F.P. Kador, J.H. Kinoshita, N.E. Sharpless. Aldose reductase inhibitors: a potential
new class of agents for the pharmacological control of certain diabetic
complications, J. Med. Chem. 28 (1985) 841-849.
8] V.F. Carvalho, E.O. Barreto, M.F. Serra, R.S. Cordeiro, M.A. Martins, Z.B. Fortes,
23

ACCEPTED MANUSCRIPT
P.M. Silva. Aldose reductase inhibitor zopolrestat restores allergic
hyporesponsiveness in alloxandiabetic rats. Eur J Pharmacol 549 (2006)
1
73-178.
[
[
9] T. Asano, Y. Saito, M. Kawakami, N.Yamada, Fidarestat (SNK-860), a potent
aldose reductase inhibitor, normalizes the elevated sorbitol accumulation in
erythrocytes of diabetic patients, J. Diabetes Complications 16 (2002) 133-138.
10] N. Hotta, Y. Akanuma, R. Kawamori, K. Matsuoka, Y. Oka, M. Shichiri, T.
Toyota, M. Nakashima, I. Yoshimura, N. Sakamoto, Y. Shigeta, Long-term
clinical effects of epalrestat, an aldose reductase inhibitor, on diabetic peripheral
neuropathy, Diabetes Care 29 (2006) 1538-1544.
[
11] T. Matsumoto, Y. Ono, M. Kurono, A. Kuromiya, K. Nakamura, V. Bril,
Ranirestat (AS-3201), a potent aldose reductase inhibitor, reduces sorbinil levels
and improves motor nerve conduction velocity in streptozotocin-diabetes rats, J.
Pharmacol. Sci. 107 (2008) 231-237.
[
[
12] K.E. Schemmel, R.S. Padiyara, J.J. D’Souza, Aldose reductase inhibitors in the
treatment of diabetic peripheral neuropathy: A review, J. Diabetes Complications
2
4 (2010) 354-360.
13] Y. Zou, X. Qin, X. Hao, W. Zhang, S. Yang, Y. Yang, Z. Han, B. Ma, C. Zhu,
Phenolic 4-hydroxy and 3,5-dihydroxy derivatives of
-phenoxyquinoxalin-2(1H)-one as potent aldose reductase inhibitors with
antioxidant activity, Bioorg. Med. Chem. Lett. 25 (2015) 3924-3927.
14] H. Pan, Y. Wang, M. Li, G. Xu, Y. Peng, J. He, Y. Zhao, Y. Li, Q. Zhao,
3
[
24

ACCEPTED MANUSCRIPT
Terpenoids from Salvia trijuga, J. Nat. Prod. 73 (2010) 1146-1150.
15] L.M. Kutrzeba, D. Ferreira, J.K. Zjawiony, Salvinorins J from Salvia divinorum:
mutarotation in the neoclerodane system, J. Nat. Prod. 72 (2009) 1361-1363.
16] M.I. Choudhary, A. Hussain, Z. Ali, A. Adhikari, S.A. Sattar, S.A.M. Ayatollahi,
A.M.A. Al-Majid, A. Rahman, Diterpenoids including a novel dimeric conjugate
from Salvia leriaefolia, Planta Med. 78 (2012) 269-275.
[
[
[
17] R. Ejtahed, T. Radjabian, S. Tafreshi, Expression analysis of phenylalanine
ammonia lyase gene and rosmarinic acid production in Salvia officinalis and
Salvia virgata shoots under salicylic acid elicitation, Appl. Biochem. Biotechnol
1
76 (2015) 1846-1858.
[
[
18] Y. Lu, Y.L. Foo, Polyphenolics of Salvia, Phytochemistry 59 (2002) 117-140.
19] M. Petersen, M.S. Simmonds, Rosmarinic acid, Phytochemistry 62 (2003)
1
21-125.
[
[
20] M. Petersen, Rosmarinic acid: new aspects, Phytochem. Rev. 12 (2013) 207-227.
21] E.L. Bakota, J.K. Winkler-Moser, M.A. Berhow, F.J. Eller, S.F. Vaughn,
Antioxidant activity and sensory evaluation of a rosmarinic acid-enriched extract
of Salvia officinalis, J. Food Sci. 80 (2015) C711-C717.
[
22] B. Huang, B. Yi, Y. Duan, L. Sun, X. Yu, J. Guo, W. Chen, Characterization and
expression profiling of tyrosine aminotransferase gene from Salvia miltiorrhiza
(
Dan-shen) in rosmarinic acid biosynthesis pathway, Mol. Biol. Rep. 35 (2008)
01-612.
23] N. Martins, L. Barros, C. Santos-Buelga, M. Henriques, S. Silva, I. Ferreira,
6
[
25

ACCEPTED MANUSCRIPT
Evaluation of bioactive properties and phenolic compounds in different extracts
prepared from Salvia officinalis L., Food Chem. 170 (2015) 378-385.
[
[
24] V. Swarup, J. Ghosh, S. Ghosh, A. Saxena, A. Basu, Anti viral and anti
inflammatory effect of rosmarinic acid in an experimental murine model of
Japanese encephalitis, Antimicrob. Agents Chemother. 51 (2007) 3367-3370.
25] T. Tanaka, T. Nakashima, T. Ueda, K. Tomii, I. Kouno, Facile discrimination of
aldose enantiomers by reversed-phase HPLC, Chem. Pharm. Bull. 55 (2007)
8
99-901.
[
[
26] M. Xie, Z. Shen, Inhibition of aldose reductase from rat lens by flavonoids, Acta
Pharmacol. Sin. 10 (1986) 721-724.
27] E. Claire, S. Schwaiger, B. Banaigs, H. Stuppner, F. Gafner, Distribution of a
new rosmarinic acid derivative in Eryngium alpinum L. and other Apiaceae, J.
Agric. Food Chem. 53 (2005) 4367-4372.
[
[
[
[
28] H. Huang, Phenolic compounds of Isodon oresbius, J. Nat. Prod. 59 (1996)
1
079-1080.
29] H. Luo, B.J. Wu, M. Wu, Z. Yong, M. Niwa, Y. Hirata, Pigments from Salvia
Miltiorrhiza, Phytochemistry 24 (1985) 815-817.
30] G. Wang, T. Li, F. Deng, Y. Li, W. Ye, Five new phenolic glycosides from
Hedyotis scandens, Bioorg. Med. Chem. Lett. 23 (2013) 1379-1382.
31] P. Alexiou, V.J. Demopoulos,
A
diverse series of substituted
benzenesulfonamides as aldose reductase inhibitors with antioxidant activity:
design, synthesis, and in vitro activity, J. Med. Chem. 53 (2010) 7756-7766.
26

ACCEPTED MANUSCRIPT
[
32] T. Ha, J. Lee, M. Lee, B. Lee, H. Kwon, C. Park, K. Shim, H. Kim, I. Baek, D.
Jang, Isolation and identification of phenolic compounds from the seeds of
Perilla frutescens (L.) and their inhibitory activities against a-glucosidase and
aldose reductase, Food Chem. 135 (2012) 1397-1403.
[
[
33] H. Li, S. Hwang, B. Kang, J. Hong, S. Lim, Inhibitory effects of Colocasia
esculenta (L.) Schott constituents on aldose reductase, Molecules 19 (2014)
1
3212-13224.
34] R. Kasimu, K. Tanaka, Y. Tezuka, Z. Gong, J. Li, P. Basnet, T. Namba, S. Kadota,
Comparative study of seventeen Salvia plants: Aldose reductase inhibitory
activity of water and MeOH extracts and liquid chromatagraph-mass
spectrometry (LC-MS) analysis of water extracts, Chem. Pharm. Bull. 46 (1998)
5
00-504.
27

ACCEPTED MANUSCRIPT
Figure Legends
Figure 1. The structures of seven compounds (1–7) isolated from S. grandifolia and
ten derivatives (8–17) of rosmarinic acid (3).
Figure 2. Key HMBC correlations (
) and NOEs (
) of compounds 1 and 2.
Figure 3. Structures of 1-A and 1-B (a), experimental ECD spectra of 1 in MeOH and
calculated ECD spectra of 1-A and 1-B (b).
Figure 4. Synthetic routes of compounds 8–17.
Figure 5. The AR inhibitory values of seven test compounds and positive control
epalrestat (A: 1, 3, 4, 5; B: 9, 10, 11).
28

ACCEPTED MANUSCRIPT
Fig. 1
29